From: "Raju P.L.S.S.S.N" <[email protected]>
tcs->lock was introduced to serialize access with in TCS group. But
even without tcs->lock, drv->lock is serving the same purpose. So
use a single drv->lock.
Other optimizations include -
- Remove locking around clear_bit() in IRQ handler. clear_bit() is
atomic.
- Remove redundant read of TCS registers.
- Use spin_lock instead of _irq variants as the locks are not held
in interrupt context.
Fixes: 658628 ("drivers: qcom: rpmh-rsc: add RPMH controller for QCOM
SoCs")
Signed-off-by: Raju P.L.S.S.S.N <[email protected]>
Signed-off-by: Lina Iyer <[email protected]>
---
drivers/soc/qcom/rpmh-internal.h | 2 --
drivers/soc/qcom/rpmh-rsc.c | 37 +++++++++++---------------------
drivers/soc/qcom/rpmh.c | 20 +++++++----------
3 files changed, 21 insertions(+), 38 deletions(-)
diff --git a/drivers/soc/qcom/rpmh-internal.h b/drivers/soc/qcom/rpmh-internal.h
index a7bbbb67991c..969d5030860e 100644
--- a/drivers/soc/qcom/rpmh-internal.h
+++ b/drivers/soc/qcom/rpmh-internal.h
@@ -28,7 +28,6 @@ struct rsc_drv;
* @offset: start of the TCS group relative to the TCSes in the RSC
* @num_tcs: number of TCSes in this type
* @ncpt: number of commands in each TCS
- * @lock: lock for synchronizing this TCS writes
* @req: requests that are sent from the TCS
* @cmd_cache: flattened cache of cmds in sleep/wake TCS
* @slots: indicates which of @cmd_addr are occupied
@@ -40,7 +39,6 @@ struct tcs_group {
u32 offset;
int num_tcs;
int ncpt;
- spinlock_t lock;
const struct tcs_request *req[MAX_TCS_PER_TYPE];
u32 *cmd_cache;
DECLARE_BITMAP(slots, MAX_TCS_SLOTS);
diff --git a/drivers/soc/qcom/rpmh-rsc.c b/drivers/soc/qcom/rpmh-rsc.c
index e278fc11fe5c..92461311aef3 100644
--- a/drivers/soc/qcom/rpmh-rsc.c
+++ b/drivers/soc/qcom/rpmh-rsc.c
@@ -93,8 +93,7 @@ static void write_tcs_reg_sync(struct rsc_drv *drv, int reg, int tcs_id,
static bool tcs_is_free(struct rsc_drv *drv, int tcs_id)
{
- return !test_bit(tcs_id, drv->tcs_in_use) &&
- read_tcs_reg(drv, RSC_DRV_STATUS, tcs_id, 0);
+ return !test_bit(tcs_id, drv->tcs_in_use);
}
static struct tcs_group *get_tcs_of_type(struct rsc_drv *drv, int type)
@@ -104,29 +103,28 @@ static struct tcs_group *get_tcs_of_type(struct rsc_drv *drv, int type)
static int tcs_invalidate(struct rsc_drv *drv, int type)
{
- int m;
+ int m, ret = 0;
struct tcs_group *tcs;
tcs = get_tcs_of_type(drv, type);
- spin_lock(&tcs->lock);
- if (bitmap_empty(tcs->slots, MAX_TCS_SLOTS)) {
- spin_unlock(&tcs->lock);
- return 0;
- }
+ spin_lock(&drv->lock);
+ if (bitmap_empty(tcs->slots, MAX_TCS_SLOTS))
+ goto done;
for (m = tcs->offset; m < tcs->offset + tcs->num_tcs; m++) {
if (!tcs_is_free(drv, m)) {
- spin_unlock(&tcs->lock);
- return -EAGAIN;
+ ret = -EAGAIN;
+ goto done;
}
write_tcs_reg_sync(drv, RSC_DRV_CMD_ENABLE, m, 0);
write_tcs_reg_sync(drv, RSC_DRV_CMD_WAIT_FOR_CMPL, m, 0);
}
bitmap_zero(tcs->slots, MAX_TCS_SLOTS);
- spin_unlock(&tcs->lock);
- return 0;
+done:
+ spin_unlock(&drv->lock);
+ return ret;
}
/**
@@ -242,9 +240,7 @@ static irqreturn_t tcs_tx_done(int irq, void *p)
write_tcs_reg(drv, RSC_DRV_CMD_ENABLE, i, 0);
write_tcs_reg(drv, RSC_DRV_CMD_WAIT_FOR_CMPL, i, 0);
write_tcs_reg(drv, RSC_DRV_IRQ_CLEAR, 0, BIT(i));
- spin_lock(&drv->lock);
clear_bit(i, drv->tcs_in_use);
- spin_unlock(&drv->lock);
if (req)
rpmh_tx_done(req, err);
}
@@ -349,14 +345,12 @@ static int tcs_write(struct rsc_drv *drv, const struct tcs_request *msg)
{
struct tcs_group *tcs;
int tcs_id;
- unsigned long flags;
int ret;
tcs = get_tcs_for_msg(drv, msg);
if (IS_ERR(tcs))
return PTR_ERR(tcs);
- spin_lock_irqsave(&tcs->lock, flags);
spin_lock(&drv->lock);
/*
* The h/w does not like if we send a request to the same address,
@@ -364,26 +358,23 @@ static int tcs_write(struct rsc_drv *drv, const struct tcs_request *msg)
*/
ret = check_for_req_inflight(drv, tcs, msg);
if (ret) {
- spin_unlock(&drv->lock);
goto done_write;
}
tcs_id = find_free_tcs(tcs);
if (tcs_id < 0) {
ret = tcs_id;
- spin_unlock(&drv->lock);
goto done_write;
}
tcs->req[tcs_id - tcs->offset] = msg;
set_bit(tcs_id, drv->tcs_in_use);
- spin_unlock(&drv->lock);
__tcs_buffer_write(drv, tcs_id, 0, msg);
__tcs_trigger(drv, tcs_id);
done_write:
- spin_unlock_irqrestore(&tcs->lock, flags);
+ spin_unlock(&drv->lock);
return ret;
}
@@ -481,19 +472,18 @@ static int tcs_ctrl_write(struct rsc_drv *drv, const struct tcs_request *msg)
{
struct tcs_group *tcs;
int tcs_id = 0, cmd_id = 0;
- unsigned long flags;
int ret;
tcs = get_tcs_for_msg(drv, msg);
if (IS_ERR(tcs))
return PTR_ERR(tcs);
- spin_lock_irqsave(&tcs->lock, flags);
+ spin_lock(&drv->lock);
/* find the TCS id and the command in the TCS to write to */
ret = find_slots(tcs, msg, &tcs_id, &cmd_id);
if (!ret)
__tcs_buffer_write(drv, tcs_id, cmd_id, msg);
- spin_unlock_irqrestore(&tcs->lock, flags);
+ spin_unlock(&drv->lock);
return ret;
}
@@ -584,7 +574,6 @@ static int rpmh_probe_tcs_config(struct platform_device *pdev,
tcs->type = tcs_cfg[i].type;
tcs->num_tcs = tcs_cfg[i].n;
tcs->ncpt = ncpt;
- spin_lock_init(&tcs->lock);
if (!tcs->num_tcs || tcs->type == CONTROL_TCS)
continue;
diff --git a/drivers/soc/qcom/rpmh.c b/drivers/soc/qcom/rpmh.c
index 035091fd44b8..12f830610b94 100644
--- a/drivers/soc/qcom/rpmh.c
+++ b/drivers/soc/qcom/rpmh.c
@@ -118,9 +118,8 @@ static struct cache_req *cache_rpm_request(struct rpmh_ctrlr *ctrlr,
struct tcs_cmd *cmd)
{
struct cache_req *req;
- unsigned long flags;
- spin_lock_irqsave(&ctrlr->cache_lock, flags);
+ spin_lock(&ctrlr->cache_lock);
req = __find_req(ctrlr, cmd->addr);
if (req)
goto existing;
@@ -154,7 +153,7 @@ static struct cache_req *cache_rpm_request(struct rpmh_ctrlr *ctrlr,
ctrlr->dirty = true;
unlock:
- spin_unlock_irqrestore(&ctrlr->cache_lock, flags);
+ spin_unlock(&ctrlr->cache_lock);
return req;
}
@@ -283,23 +282,21 @@ EXPORT_SYMBOL(rpmh_write);
static void cache_batch(struct rpmh_ctrlr *ctrlr, struct batch_cache_req *req)
{
- unsigned long flags;
- spin_lock_irqsave(&ctrlr->cache_lock, flags);
+ spin_lock(&ctrlr->cache_lock);
list_add_tail(&req->list, &ctrlr->batch_cache);
- spin_unlock_irqrestore(&ctrlr->cache_lock, flags);
+ spin_unlock(&ctrlr->cache_lock);
}
static int flush_batch(struct rpmh_ctrlr *ctrlr)
{
struct batch_cache_req *req;
const struct rpmh_request *rpm_msg;
- unsigned long flags;
int ret = 0;
int i;
/* Send Sleep/Wake requests to the controller, expect no response */
- spin_lock_irqsave(&ctrlr->cache_lock, flags);
+ spin_lock(&ctrlr->cache_lock);
list_for_each_entry(req, &ctrlr->batch_cache, list) {
for (i = 0; i < req->count; i++) {
rpm_msg = req->rpm_msgs + i;
@@ -309,7 +306,7 @@ static int flush_batch(struct rpmh_ctrlr *ctrlr)
break;
}
}
- spin_unlock_irqrestore(&ctrlr->cache_lock, flags);
+ spin_unlock(&ctrlr->cache_lock);
return ret;
}
@@ -317,13 +314,12 @@ static int flush_batch(struct rpmh_ctrlr *ctrlr)
static void invalidate_batch(struct rpmh_ctrlr *ctrlr)
{
struct batch_cache_req *req, *tmp;
- unsigned long flags;
- spin_lock_irqsave(&ctrlr->cache_lock, flags);
+ spin_lock(&ctrlr->cache_lock);
list_for_each_entry_safe(req, tmp, &ctrlr->batch_cache, list)
kfree(req);
INIT_LIST_HEAD(&ctrlr->batch_cache);
- spin_unlock_irqrestore(&ctrlr->cache_lock, flags);
+ spin_unlock(&ctrlr->cache_lock);
}
/**
--
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project
When triggering a TCS to send its contents, reading back the trigger
value may return an incorrect value. That is because, writing the
trigger may raise an interrupt which could be handled immediately and
the trigger value could be reset in the interrupt handler. By doing a
read back we may end up spinning waiting for the value we wrote.
Fixes: 658628 ("drivers: qcom: rpmh-rsc: add RPMH controller for QCOM
SoCs")
Signed-off-by: Lina Iyer <[email protected]>
---
drivers/soc/qcom/rpmh-rsc.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/soc/qcom/rpmh-rsc.c b/drivers/soc/qcom/rpmh-rsc.c
index 92461311aef3..2fc2fa879480 100644
--- a/drivers/soc/qcom/rpmh-rsc.c
+++ b/drivers/soc/qcom/rpmh-rsc.c
@@ -300,7 +300,7 @@ static void __tcs_trigger(struct rsc_drv *drv, int tcs_id)
enable = TCS_AMC_MODE_ENABLE;
write_tcs_reg_sync(drv, RSC_DRV_CONTROL, tcs_id, enable);
enable |= TCS_AMC_MODE_TRIGGER;
- write_tcs_reg_sync(drv, RSC_DRV_CONTROL, tcs_id, enable);
+ write_tcs_reg(drv, RSC_DRV_CONTROL, tcs_id, enable);
}
static int check_for_req_inflight(struct rsc_drv *drv, struct tcs_group *tcs,
--
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project
Switching Andy's email address.
On Mon, Jul 01 2019 at 09:32 -0600, Lina Iyer wrote:
>From: "Raju P.L.S.S.S.N" <[email protected]>
>
>tcs->lock was introduced to serialize access with in TCS group. But
>even without tcs->lock, drv->lock is serving the same purpose. So
>use a single drv->lock.
>
>Other optimizations include -
> - Remove locking around clear_bit() in IRQ handler. clear_bit() is
> atomic.
> - Remove redundant read of TCS registers.
> - Use spin_lock instead of _irq variants as the locks are not held
> in interrupt context.
>
>Fixes: 658628 ("drivers: qcom: rpmh-rsc: add RPMH controller for QCOM
>SoCs")
>Signed-off-by: Raju P.L.S.S.S.N <[email protected]>
>Signed-off-by: Lina Iyer <[email protected]>
>---
> drivers/soc/qcom/rpmh-internal.h | 2 --
> drivers/soc/qcom/rpmh-rsc.c | 37 +++++++++++---------------------
> drivers/soc/qcom/rpmh.c | 20 +++++++----------
> 3 files changed, 21 insertions(+), 38 deletions(-)
>
>diff --git a/drivers/soc/qcom/rpmh-internal.h b/drivers/soc/qcom/rpmh-internal.h
>index a7bbbb67991c..969d5030860e 100644
>--- a/drivers/soc/qcom/rpmh-internal.h
>+++ b/drivers/soc/qcom/rpmh-internal.h
>@@ -28,7 +28,6 @@ struct rsc_drv;
> * @offset: start of the TCS group relative to the TCSes in the RSC
> * @num_tcs: number of TCSes in this type
> * @ncpt: number of commands in each TCS
>- * @lock: lock for synchronizing this TCS writes
> * @req: requests that are sent from the TCS
> * @cmd_cache: flattened cache of cmds in sleep/wake TCS
> * @slots: indicates which of @cmd_addr are occupied
>@@ -40,7 +39,6 @@ struct tcs_group {
> u32 offset;
> int num_tcs;
> int ncpt;
>- spinlock_t lock;
> const struct tcs_request *req[MAX_TCS_PER_TYPE];
> u32 *cmd_cache;
> DECLARE_BITMAP(slots, MAX_TCS_SLOTS);
>diff --git a/drivers/soc/qcom/rpmh-rsc.c b/drivers/soc/qcom/rpmh-rsc.c
>index e278fc11fe5c..92461311aef3 100644
>--- a/drivers/soc/qcom/rpmh-rsc.c
>+++ b/drivers/soc/qcom/rpmh-rsc.c
>@@ -93,8 +93,7 @@ static void write_tcs_reg_sync(struct rsc_drv *drv, int reg, int tcs_id,
>
> static bool tcs_is_free(struct rsc_drv *drv, int tcs_id)
> {
>- return !test_bit(tcs_id, drv->tcs_in_use) &&
>- read_tcs_reg(drv, RSC_DRV_STATUS, tcs_id, 0);
>+ return !test_bit(tcs_id, drv->tcs_in_use);
> }
>
> static struct tcs_group *get_tcs_of_type(struct rsc_drv *drv, int type)
>@@ -104,29 +103,28 @@ static struct tcs_group *get_tcs_of_type(struct rsc_drv *drv, int type)
>
> static int tcs_invalidate(struct rsc_drv *drv, int type)
> {
>- int m;
>+ int m, ret = 0;
> struct tcs_group *tcs;
>
> tcs = get_tcs_of_type(drv, type);
>
>- spin_lock(&tcs->lock);
>- if (bitmap_empty(tcs->slots, MAX_TCS_SLOTS)) {
>- spin_unlock(&tcs->lock);
>- return 0;
>- }
>+ spin_lock(&drv->lock);
>+ if (bitmap_empty(tcs->slots, MAX_TCS_SLOTS))
>+ goto done;
>
> for (m = tcs->offset; m < tcs->offset + tcs->num_tcs; m++) {
> if (!tcs_is_free(drv, m)) {
>- spin_unlock(&tcs->lock);
>- return -EAGAIN;
>+ ret = -EAGAIN;
>+ goto done;
> }
> write_tcs_reg_sync(drv, RSC_DRV_CMD_ENABLE, m, 0);
> write_tcs_reg_sync(drv, RSC_DRV_CMD_WAIT_FOR_CMPL, m, 0);
> }
> bitmap_zero(tcs->slots, MAX_TCS_SLOTS);
>- spin_unlock(&tcs->lock);
>
>- return 0;
>+done:
>+ spin_unlock(&drv->lock);
>+ return ret;
> }
>
> /**
>@@ -242,9 +240,7 @@ static irqreturn_t tcs_tx_done(int irq, void *p)
> write_tcs_reg(drv, RSC_DRV_CMD_ENABLE, i, 0);
> write_tcs_reg(drv, RSC_DRV_CMD_WAIT_FOR_CMPL, i, 0);
> write_tcs_reg(drv, RSC_DRV_IRQ_CLEAR, 0, BIT(i));
>- spin_lock(&drv->lock);
> clear_bit(i, drv->tcs_in_use);
>- spin_unlock(&drv->lock);
> if (req)
> rpmh_tx_done(req, err);
> }
>@@ -349,14 +345,12 @@ static int tcs_write(struct rsc_drv *drv, const struct tcs_request *msg)
> {
> struct tcs_group *tcs;
> int tcs_id;
>- unsigned long flags;
> int ret;
>
> tcs = get_tcs_for_msg(drv, msg);
> if (IS_ERR(tcs))
> return PTR_ERR(tcs);
>
>- spin_lock_irqsave(&tcs->lock, flags);
> spin_lock(&drv->lock);
> /*
> * The h/w does not like if we send a request to the same address,
>@@ -364,26 +358,23 @@ static int tcs_write(struct rsc_drv *drv, const struct tcs_request *msg)
> */
> ret = check_for_req_inflight(drv, tcs, msg);
> if (ret) {
>- spin_unlock(&drv->lock);
> goto done_write;
> }
>
> tcs_id = find_free_tcs(tcs);
> if (tcs_id < 0) {
> ret = tcs_id;
>- spin_unlock(&drv->lock);
> goto done_write;
> }
>
> tcs->req[tcs_id - tcs->offset] = msg;
> set_bit(tcs_id, drv->tcs_in_use);
>- spin_unlock(&drv->lock);
>
> __tcs_buffer_write(drv, tcs_id, 0, msg);
> __tcs_trigger(drv, tcs_id);
>
> done_write:
>- spin_unlock_irqrestore(&tcs->lock, flags);
>+ spin_unlock(&drv->lock);
> return ret;
> }
>
>@@ -481,19 +472,18 @@ static int tcs_ctrl_write(struct rsc_drv *drv, const struct tcs_request *msg)
> {
> struct tcs_group *tcs;
> int tcs_id = 0, cmd_id = 0;
>- unsigned long flags;
> int ret;
>
> tcs = get_tcs_for_msg(drv, msg);
> if (IS_ERR(tcs))
> return PTR_ERR(tcs);
>
>- spin_lock_irqsave(&tcs->lock, flags);
>+ spin_lock(&drv->lock);
> /* find the TCS id and the command in the TCS to write to */
> ret = find_slots(tcs, msg, &tcs_id, &cmd_id);
> if (!ret)
> __tcs_buffer_write(drv, tcs_id, cmd_id, msg);
>- spin_unlock_irqrestore(&tcs->lock, flags);
>+ spin_unlock(&drv->lock);
>
> return ret;
> }
>@@ -584,7 +574,6 @@ static int rpmh_probe_tcs_config(struct platform_device *pdev,
> tcs->type = tcs_cfg[i].type;
> tcs->num_tcs = tcs_cfg[i].n;
> tcs->ncpt = ncpt;
>- spin_lock_init(&tcs->lock);
>
> if (!tcs->num_tcs || tcs->type == CONTROL_TCS)
> continue;
>diff --git a/drivers/soc/qcom/rpmh.c b/drivers/soc/qcom/rpmh.c
>index 035091fd44b8..12f830610b94 100644
>--- a/drivers/soc/qcom/rpmh.c
>+++ b/drivers/soc/qcom/rpmh.c
>@@ -118,9 +118,8 @@ static struct cache_req *cache_rpm_request(struct rpmh_ctrlr *ctrlr,
> struct tcs_cmd *cmd)
> {
> struct cache_req *req;
>- unsigned long flags;
>
>- spin_lock_irqsave(&ctrlr->cache_lock, flags);
>+ spin_lock(&ctrlr->cache_lock);
> req = __find_req(ctrlr, cmd->addr);
> if (req)
> goto existing;
>@@ -154,7 +153,7 @@ static struct cache_req *cache_rpm_request(struct rpmh_ctrlr *ctrlr,
>
> ctrlr->dirty = true;
> unlock:
>- spin_unlock_irqrestore(&ctrlr->cache_lock, flags);
>+ spin_unlock(&ctrlr->cache_lock);
>
> return req;
> }
>@@ -283,23 +282,21 @@ EXPORT_SYMBOL(rpmh_write);
>
> static void cache_batch(struct rpmh_ctrlr *ctrlr, struct batch_cache_req *req)
> {
>- unsigned long flags;
>
>- spin_lock_irqsave(&ctrlr->cache_lock, flags);
>+ spin_lock(&ctrlr->cache_lock);
> list_add_tail(&req->list, &ctrlr->batch_cache);
>- spin_unlock_irqrestore(&ctrlr->cache_lock, flags);
>+ spin_unlock(&ctrlr->cache_lock);
> }
>
> static int flush_batch(struct rpmh_ctrlr *ctrlr)
> {
> struct batch_cache_req *req;
> const struct rpmh_request *rpm_msg;
>- unsigned long flags;
> int ret = 0;
> int i;
>
> /* Send Sleep/Wake requests to the controller, expect no response */
>- spin_lock_irqsave(&ctrlr->cache_lock, flags);
>+ spin_lock(&ctrlr->cache_lock);
> list_for_each_entry(req, &ctrlr->batch_cache, list) {
> for (i = 0; i < req->count; i++) {
> rpm_msg = req->rpm_msgs + i;
>@@ -309,7 +306,7 @@ static int flush_batch(struct rpmh_ctrlr *ctrlr)
> break;
> }
> }
>- spin_unlock_irqrestore(&ctrlr->cache_lock, flags);
>+ spin_unlock(&ctrlr->cache_lock);
>
> return ret;
> }
>@@ -317,13 +314,12 @@ static int flush_batch(struct rpmh_ctrlr *ctrlr)
> static void invalidate_batch(struct rpmh_ctrlr *ctrlr)
> {
> struct batch_cache_req *req, *tmp;
>- unsigned long flags;
>
>- spin_lock_irqsave(&ctrlr->cache_lock, flags);
>+ spin_lock(&ctrlr->cache_lock);
> list_for_each_entry_safe(req, tmp, &ctrlr->batch_cache, list)
> kfree(req);
> INIT_LIST_HEAD(&ctrlr->batch_cache);
>- spin_unlock_irqrestore(&ctrlr->cache_lock, flags);
>+ spin_unlock(&ctrlr->cache_lock);
> }
>
> /**
>--
>The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
>a Linux Foundation Collaborative Project
>
Switching Andy's email address.
On Mon, Jul 01 2019 at 09:32 -0600, Lina Iyer wrote:
>When triggering a TCS to send its contents, reading back the trigger
>value may return an incorrect value. That is because, writing the
>trigger may raise an interrupt which could be handled immediately and
>the trigger value could be reset in the interrupt handler. By doing a
>read back we may end up spinning waiting for the value we wrote.
>
>Fixes: 658628 ("drivers: qcom: rpmh-rsc: add RPMH controller for QCOM
>SoCs")
>Signed-off-by: Lina Iyer <[email protected]>
>---
> drivers/soc/qcom/rpmh-rsc.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
>diff --git a/drivers/soc/qcom/rpmh-rsc.c b/drivers/soc/qcom/rpmh-rsc.c
>index 92461311aef3..2fc2fa879480 100644
>--- a/drivers/soc/qcom/rpmh-rsc.c
>+++ b/drivers/soc/qcom/rpmh-rsc.c
>@@ -300,7 +300,7 @@ static void __tcs_trigger(struct rsc_drv *drv, int tcs_id)
> enable = TCS_AMC_MODE_ENABLE;
> write_tcs_reg_sync(drv, RSC_DRV_CONTROL, tcs_id, enable);
> enable |= TCS_AMC_MODE_TRIGGER;
>- write_tcs_reg_sync(drv, RSC_DRV_CONTROL, tcs_id, enable);
>+ write_tcs_reg(drv, RSC_DRV_CONTROL, tcs_id, enable);
> }
>
> static int check_for_req_inflight(struct rsc_drv *drv, struct tcs_group *tcs,
>--
>The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
>a Linux Foundation Collaborative Project
>
Quoting Lina Iyer (2019-07-01 08:29:06)
> From: "Raju P.L.S.S.S.N" <[email protected]>
>
> tcs->lock was introduced to serialize access with in TCS group. But
> even without tcs->lock, drv->lock is serving the same purpose. So
> use a single drv->lock.
Isn't the downside now that we're going to be serializing access to the
different TCSes when two are being written in parallel or waited on? I
thought that was the whole point of splitting the lock into a TCS lock
and a general "driver" lock that protects the global driver state vs.
the specific TCS state.
>
> Other optimizations include -
> - Remove locking around clear_bit() in IRQ handler. clear_bit() is
> atomic.
> - Remove redundant read of TCS registers.
> - Use spin_lock instead of _irq variants as the locks are not held
> in interrupt context.
Can you please split this patch up into 3 or 4 different patches? I'm
not sure why any of these patches are marked with Fixes either. It's an
optimization patch, not a fix patch, unless the optimization is really
large somehow?
>
> Fixes: 658628 ("drivers: qcom: rpmh-rsc: add RPMH controller for QCOM
> SoCs")
> Signed-off-by: Raju P.L.S.S.S.N <[email protected]>
> Signed-off-by: Lina Iyer <[email protected]>
> ---
> drivers/soc/qcom/rpmh-internal.h | 2 --
> drivers/soc/qcom/rpmh-rsc.c | 37 +++++++++++---------------------
> drivers/soc/qcom/rpmh.c | 20 +++++++----------
> 3 files changed, 21 insertions(+), 38 deletions(-)
>
> diff --git a/drivers/soc/qcom/rpmh-internal.h b/drivers/soc/qcom/rpmh-internal.h
> index a7bbbb67991c..969d5030860e 100644
> --- a/drivers/soc/qcom/rpmh-internal.h
> +++ b/drivers/soc/qcom/rpmh-internal.h
> diff --git a/drivers/soc/qcom/rpmh-rsc.c b/drivers/soc/qcom/rpmh-rsc.c
> index e278fc11fe5c..92461311aef3 100644
> --- a/drivers/soc/qcom/rpmh-rsc.c
> +++ b/drivers/soc/qcom/rpmh-rsc.c
> @@ -93,8 +93,7 @@ static void write_tcs_reg_sync(struct rsc_drv *drv, int reg, int tcs_id,
>
> static bool tcs_is_free(struct rsc_drv *drv, int tcs_id)
> {
> - return !test_bit(tcs_id, drv->tcs_in_use) &&
> - read_tcs_reg(drv, RSC_DRV_STATUS, tcs_id, 0);
> + return !test_bit(tcs_id, drv->tcs_in_use);
This can be a different patch. Why is reading the tcs register
redundant? Please put that information in the commit text.
Quoting Lina Iyer (2019-07-01 08:29:07)
> When triggering a TCS to send its contents, reading back the trigger
> value may return an incorrect value. That is because, writing the
> trigger may raise an interrupt which could be handled immediately and
> the trigger value could be reset in the interrupt handler. By doing a
> read back we may end up spinning waiting for the value we wrote.
Doesn't this need to be squashed into the patch that gets rid of the
irqs disabled state of this code? It sounds an awful lot like this
problem only happens now because the previous patch removed the
irqsave/irqrestore code around this function.
On Fri, Jul 19 2019 at 12:22 -0600, Stephen Boyd wrote:
>Quoting Lina Iyer (2019-07-01 08:29:07)
>> When triggering a TCS to send its contents, reading back the trigger
>> value may return an incorrect value. That is because, writing the
>> trigger may raise an interrupt which could be handled immediately and
>> the trigger value could be reset in the interrupt handler. By doing a
>> read back we may end up spinning waiting for the value we wrote.
>
>Doesn't this need to be squashed into the patch that gets rid of the
>irqs disabled state of this code? It sounds an awful lot like this
>problem only happens now because the previous patch removed the
>irqsave/irqrestore code around this function.
>
True. Could be rolled into that fix.
--Lina
On Fri, Jul 19 2019 at 12:20 -0600, Stephen Boyd wrote:
>Quoting Lina Iyer (2019-07-01 08:29:06)
>> From: "Raju P.L.S.S.S.N" <[email protected]>
>>
>> tcs->lock was introduced to serialize access with in TCS group. But
>> even without tcs->lock, drv->lock is serving the same purpose. So
>> use a single drv->lock.
>
>Isn't the downside now that we're going to be serializing access to the
>different TCSes when two are being written in parallel or waited on? I
>thought that was the whole point of splitting the lock into a TCS lock
>and a general "driver" lock that protects the global driver state vs.
>the specific TCS state.
>
Yes but we were holding the drv->lock as well as tcs->lock for the most
critical of the path anyways (writing to TCS). The added complexity
doesn't seem to help reduce the latency that it expected to reduce.
>>
>> Other optimizations include -
>> - Remove locking around clear_bit() in IRQ handler. clear_bit() is
>> atomic.
>> - Remove redundant read of TCS registers.
>> - Use spin_lock instead of _irq variants as the locks are not held
>> in interrupt context.
>
>Can you please split this patch up into 3 or 4 different patches? I'm
>not sure why any of these patches are marked with Fixes either. It's an
>optimization patch, not a fix patch, unless the optimization is really
>large somehow?
>
Okay. I will try that.
>>
>> Fixes: 658628 ("drivers: qcom: rpmh-rsc: add RPMH controller for QCOM
>> SoCs")
>> Signed-off-by: Raju P.L.S.S.S.N <[email protected]>
>> Signed-off-by: Lina Iyer <[email protected]>
>> ---
>> drivers/soc/qcom/rpmh-internal.h | 2 --
>> drivers/soc/qcom/rpmh-rsc.c | 37 +++++++++++---------------------
>> drivers/soc/qcom/rpmh.c | 20 +++++++----------
>> 3 files changed, 21 insertions(+), 38 deletions(-)
>>
>> diff --git a/drivers/soc/qcom/rpmh-internal.h b/drivers/soc/qcom/rpmh-internal.h
>> index a7bbbb67991c..969d5030860e 100644
>> --- a/drivers/soc/qcom/rpmh-internal.h
>> +++ b/drivers/soc/qcom/rpmh-internal.h
>> diff --git a/drivers/soc/qcom/rpmh-rsc.c b/drivers/soc/qcom/rpmh-rsc.c
>> index e278fc11fe5c..92461311aef3 100644
>> --- a/drivers/soc/qcom/rpmh-rsc.c
>> +++ b/drivers/soc/qcom/rpmh-rsc.c
>> @@ -93,8 +93,7 @@ static void write_tcs_reg_sync(struct rsc_drv *drv, int reg, int tcs_id,
>>
>> static bool tcs_is_free(struct rsc_drv *drv, int tcs_id)
>> {
>> - return !test_bit(tcs_id, drv->tcs_in_use) &&
>> - read_tcs_reg(drv, RSC_DRV_STATUS, tcs_id, 0);
>> + return !test_bit(tcs_id, drv->tcs_in_use);
>
>This can be a diffedjusted rent patch. Why is reading the tcs register
>redundant? Please put that information in the commit text.
>
The tcs_in_use, is adjusted along with the DRV_STS and reading the
tcs_in_use should be enough.
Thanks for your review Stephen.
--Lina
Quoting Lina Iyer (2019-07-22 09:20:03)
> On Fri, Jul 19 2019 at 12:20 -0600, Stephen Boyd wrote:
> >Quoting Lina Iyer (2019-07-01 08:29:06)
> >> From: "Raju P.L.S.S.S.N" <[email protected]>
> >>
> >> tcs->lock was introduced to serialize access with in TCS group. But
> >> even without tcs->lock, drv->lock is serving the same purpose. So
> >> use a single drv->lock.
> >
> >Isn't the downside now that we're going to be serializing access to the
> >different TCSes when two are being written in parallel or waited on? I
> >thought that was the whole point of splitting the lock into a TCS lock
> >and a general "driver" lock that protects the global driver state vs.
> >the specific TCS state.
> >
> Yes but we were holding the drv->lock as well as tcs->lock for the most
> critical of the path anyways (writing to TCS). The added complexity
> doesn't seem to help reduce the latency that it expected to reduce.
Ok. That sort of information should be in the commit text to explain why
it's not helping with reducing the latency or throughput of the API.
On Mon, Jul 22 2019 at 12:18 -0600, Stephen Boyd wrote:
>Quoting Lina Iyer (2019-07-22 09:20:03)
>> On Fri, Jul 19 2019 at 12:20 -0600, Stephen Boyd wrote:
>> >Quoting Lina Iyer (2019-07-01 08:29:06)
>> >> From: "Raju P.L.S.S.S.N" <[email protected]>
>> >>
>> >> tcs->lock was introduced to serialize access with in TCS group. But
>> >> even without tcs->lock, drv->lock is serving the same purpose. So
>> >> use a single drv->lock.
>> >
>> >Isn't the downside now that we're going to be serializing access to the
>> >different TCSes when two are being written in parallel or waited on? I
>> >thought that was the whole point of splitting the lock into a TCS lock
>> >and a general "driver" lock that protects the global driver state vs.
>> >the specific TCS state.
>> >
>> Yes but we were holding the drv->lock as well as tcs->lock for the most
>> critical of the path anyways (writing to TCS). The added complexity
>> doesn't seem to help reduce the latency that it expected to reduce.
>
>Ok. That sort of information should be in the commit text to explain why
>it's not helping with reducing the latency or throughput of the API.
>
Will add.
--Lina